DOCSLIB.ORG

Texture Mapping

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Texture Mapping Texture Mapping CS384G – Fall 2012 Christian Miller Surface detail Surface detail ¤ Most things have a lot of detail, and simple polygons or triangle meshes are poor approximations ¤ Modeling all that detail with simple primitives would take eons, and enormous amounts of storage ¤ Rendering all of it would take forever too ¤ We can’t just give up, so we need some way to make surfaces look more detailed than they actually are… Wallpaper Texture mapping ¤ Take an image with the surface detail on it ¤ The pixels that make up the texture are often called texels ¤ Stretch it over the surface ¤ When rendering a pixel, look up the diffuse color from the texture and use the rest of the light model as usual Mapping to a surface ¤ Accomplished through texture coordinates (u, v) ¤ The texture image has coordinates (0, 0) in the lower left corner and (1, 1) in the upper right ¤ For a mesh, have user specify the (u, v) coordinates at each vertex ¤ To render a pixel, interpolate (u, v) at the intersection point, use those texcoords to look up the right color from the texture Specifying texcoords [Muse Games] [Muse Games] Alternate ways of generating UVs ¤ You don’t necessarily need to specify texcoords on a per- vertex basis ¤ Sometimes a simpler function can do it automatically for you Texture edge modes ¤ What do you do when you get a texcoord outside [0, 1]? ¤ Adopt some convention: ¤ Loop around and start at the other side (wrapping) ¤ Reflect the image backwards (mirroring) ¤ Repeat the edge pixels (clamping) ¤ Default to some other color (bordering) Texture lookup Texture filtering ¤ Simply returning the pixel you hit in the texture(nearest neighbor) looks terrible ¤ Far away, the texture is undersampled and unrecognizable ¤ Up close, the texels look huge and blocky Magnification ¤ The image looks really blocky, since each texel covers several pixels ¤ Since we have more pixels covering the area than there are texels, we need to fake data that isn’t there ¤ Blurring the image is a good idea, since even that looks better than giant sharp-edged texels ¤ The usual fix is bilinear interpolation (bilerp) Bilinear interpolation ¤ Sample neighboring texels, blend them linearly ¤ T(a, b) = (1-∆x)(1-∆y) T(i, j) + ∆x (1-∆y) T(i+1, j) + (1-∆x)∆y T(i, j+1) + ∆x ∆y T(i+1, j+1) Bilerp results Minification ¤ One pixel on the screen can cover any number of texels ¤ Coverage area in texture space is an arbitrary shape ¤ This is an undersampling issue, which means it can be addressed with anti-aliasing methods Supersampling ¤ Since we have several texels being covered by a single pixel, the analytic method is to take an average of all texels weighted by their intersection area with the pixel ¤ This is expensive and complicated ¤ Can be approximated by sending several jittered rays through the area of the pixel and averaging them ¤ Still expensive, but not complicated ¤ Most high-quality renders do this anyway, since it smooths out jaggies on edges as well as textures ¤ Used very commonly in raytracers Mipmapping ¤ In a realtime system, you may not be able to afford taking tons of samples per pixel ¤ Instead, take the original textures and make several pre- blurred versions of them ¤ Each texture is half the size of the larger one, giving a pyramid of textures from each full image ¤ Requires 1/3rd more memory than just the original texture ¤ Then at runtime, use distance from camera and surface angle to pick a version of the texture to sample Mipmap pyramid Mipmap results Standard texture mapping Standard texture mapping Other maps ¤ So far, we’ve just been using texture maps to alter the diffuse component of the lighting model ¤ In the most general case, a texture just represents some function attached to a surface ¤ What happens if we use it to store other components of the lighting model? Specular mapping ¤ Use a texture to store the intensity of the specular term Normal mapping + = ¤ Store surface normal (relative to geometry) compressed in a texture map ¤ At runtime, look up normal in texture and add it to the usual normal that’s interpolated from the vertices ¤ Makes a huge visual difference without adding geometry Normal mapping Normal mapping Normal mapping details ¤ Normals are stored in texture by making the (r, g, b) components the (x, y, z) values of the normal ¤ Z is clearly in the normal direction from the surface, but the X and Y directions are unspecified ¤ Need to add tangent and binormal vectors to form a full coordinate system at every point Displacement mapping ¤ Normal maps don’t add any detail to the silhouette of an object, since the actual geometry is still simple ¤ You can finely subdivide the geometry and use a displacement map to offset the individual vertices Environment mapping ¤ There’s no reason a texture needs to be glued to a surface, we can compute texcoords on the fly ¤ For example, if we’re running in realtime and want reflections but can’t afford to cast rays… ¤ Store the surrounding environment in a texture map ¤ Compute reflection vectors at each vertex, use those to set texture coordinates, then the environment gets mapped onto the surface as if it was reflected Spherical environment mapping ¤ Store the environment as a picture of a perfectly reflective sphere, easy math to compute texcoords ¤ Only covers a hemisphere, reflections remain fixed relative to the camera Cube mapping ¤ Spherical environment mapping can’t get all sides of an object, so you can use texture maps on the sides of a cube instead ¤ Math is more complicated, but results are better ¤ Easier to render environment map at runtime too Cube mapping Cube mapping 3D textures It’s a big iceberg ¤ There are enormous numbers of ways that texture mapping has been used ¤ Basically every cool graphics effect you see in realtime is a texturing trick ¤ A good chunk of prerendered stuff is too .
Load more

Recommended publications
  • Cardinality-Constrained Texture Filtering
    To appear in ACM TOG 32(4). Cardinality-Constrained Texture Filtering Josiah Manson Scott Schaefer Texas A&M University Texas A&M University (a) Input (b) Exact (c) 8 Texels (d) Trilinear Figure 1: We show a comparison of Lanczos´ 2 filter approximations showing (a) the 10242 input image, with downsampled images at a resolution of 892 calculated using (b) an exact Lanczos´ filter, (c) an eight texel approximation using our method, and (d) trilinear interpolation applied to a Lanczos´ filtered mipmap. Our approximation produces an image that is nearly the same as the exact filtered image while using the same number of texels as trilinear interpolation. Abstract theorem [Shannon 1949] implies that we must use a low-pass filter to remove high-frequency data from the image prior to sampling. We present a method to create high-quality sampling filters by com- There are a variety of low-pass filters, where each filter has its own bining a prescribed number of texels from several resolutions in a set of tradeoffs. Some filters remove aliasing at the cost of overblur- mipmap. Our technique provides fine control over the number of ring the image, while others blur less but allow more aliasing. Fil- texels we read per texture sample so that we can scale quality to ters that are effective at removing aliasing without overblurring sum match a memory bandwidth budget. Our method also has a fixed over a greater number of texels, which makes them expensive to cost regardless of the filter we approximate, which makes it fea- compute. As an extreme example, the sinc filter removes all high sible to approximate higher-order filters such as a Lanczos´ 2 filter frequencies and no low frequencies, but sums over an infinite num- in real-time rendering.
    [Show full text]
  • Computer Graphics Texture Filtering & Sampling Theory
    Computer Graphics Texture Filtering Philipp Slusallek Reconstruction Filter • Simple texture mapping in a ray-tracer – Ray hits surface, e.g. a triangle – Each triangle vertex also has an arbitrary texture coordinate • Map this vertex into 2D texture space (aka. texture parameterization) – Use barycentric coordinates to map hit point into texture space • Hit point generally does not exactly hit a texture sample • Use reconstruction filter to find color for hit point Texture Space 2 Nearest Neighbor “Interpolation” v c2 c3 c0 c1 u Texture Space • How to compute the color of the pixel? – Choose the closest texture sample • Rounding of the texture coordinate in texture space • c = tex[ min( u * resU , resU – 1 ) , min( v * resV , resV – 1 ) ]; 3 Bilinear Interpolation v c2 c3 1-t t c0 c1 u st 1-s Texture Space • How to compute the color of the pixel? – Interpolate between surrounding four pixels – c = (1-t) (1-s) c0 + (1-t) s c1 + t (1-s) c2 + t s c3 4 Bilinear Interpolation v c2 c3 1-t t c0 c1 u st 1-s Texture Space • Can be done in two steps: – c = (1-t) ( (1-s) c0 + s c1 ) + t ( (1-s) c2 + s c3 ) – Horizontally: twice between left and right samples using fractional part of the texture coordinate (1-s, s): • i0 = (1-s) c0 + s c1 • i1 = (1-s) c2 + s c3 – Vertically: between two intermediate results (1-t, t): • c = (1-t) i0 + t i1 5 Filtering • Magnification (Zoom-in) – Map few texels onto many pixels – Reconstruction filter: Pixel • Nearest neighbor interpolation: – Take the nearest texel • Bilinear interpolation: – Interpolation between
    [Show full text]
  • Texture Mapping: the Basics
    CHAPTER 8 Texture Mapping: The Basics by Richard S. Wright Jr. WHAT YOU’LL LEARN IN THIS CHAPTER: How To Functions You’ll Use Load texture images glTexImage/glTexSubImage Map textures to geometry glTexCoord Change the texture environment glTexEnv Set texture mapping parameters glTexParameter Generate mipmaps gluBuildMipmaps Manage multiple textures glBindTexture In the preceding chapter, we covered in detail the groundwork for loading image data into OpenGL. Image data, unless modified by pixel zoom, generally has a one-to-one corre- spondence between a pixel in an image and a pixel on the screen. In fact, this is where we get the term pixel (picture element). In this chapter, we extend this knowledge further by applying images to three-dimensional primitives. When we apply image data to a geomet- ric primitive, we call this a texture or texture map. Figure 8.1 shows the dramatic difference that can be achieved by texture mapping geometry. The cube on the left is a lit and shaded featureless surface, whereas the cube on the right shows a richness in detail that can be reasonably achieved only with texture mapping. 304 CHAPTER 8 Texture Mapping: The Basics FIGURE 8.1 The stark contrast between textured and untextured geometry. A texture image when loaded has the same makeup and arrangement as pixmaps, but now a one-to-one correspondence seldom exists between texels (the individual picture elements in a texture) and pixels on the screen. This chapter covers the basics of loading a texture map into memory and all the ways in which it may be mapped to and applied to geomet- ric primitives.
    [Show full text]
  • Spatio-Temporal Upsampling on the GPU
    Spatio-Temporal Upsampling on the GPU Robert Herzog∗ Elmar Eisemanny Karol Myszkowskiz H.-P. Seidel MPI Informatik Saarland University / MPI / Tel´ ecom´ ParisTech MPI Informatik MPI Informatik Reference Temporally-amortized Upsampling Spatial Upsampling Our Spatio-temporal Upsampling 2 fps 15 fps (PSNR 65.6) 24 fps (PSNR 66.5) 22 fps (PSNR 67.4) Figure 1: Comparison of different upsampling schemes in a fully dynamic scene with complex shading (indirect light and ambient occlusion). Abstract reduce rendering costs, suppress aliasing, and popping artifacts be- comes more and more attractive. Pixel processing is becoming increasingly expensive for real-time Our method is driven by the observation that high quality is most applications due to the complexity of today’s shaders and high- important for static elements, thus we can accept some loss if strong resolution framebuffers. However, most shading results are spa- differences occur. This has been shown to be a good assumption, tially or temporally coherent, which allows for sparse sampling and recently exploited for shadow computations [Scherzer et al. 2007]. reuse of neighboring pixel values. This paper proposes a simple To achieve our goal, we rely on a varying sampling pattern pro- framework for spatio-temporal upsampling on modern GPUs. In ducing a low-resolution image and keep several such samples over contrast to previous work, which focuses either on temporal or spa- time. Our idea is to integrate all these samples in a unified manner. tial processing on the GPU, we exploit coherence in both. Our al- gorithm combines adaptive motion-compensated filtering over time The heart of our method is a filtering strategy that combines sam- and geometry-aware upsampling in image space.
    [Show full text]
  • Antialiasing Complex Global Illumination Effects in Path-Space
    Antialiasing Complex Global Illumination Effects in Path-space Laurent Belcour1, Ling-Qi Yan2, Ravi Ramamoorthi3, and Derek Nowrouzezahrai1 1Universite´ de Montreal,´ 2UC Berkeley, 3UC San Diego We present the first method to efficiently and accurately predict antialias- imate surface footprints from this bandlimit1. We also merge ing footprints to pre-filter color-, normal-, and displacement-mapped ap- two independent unidirectional frequency analyses at path vertex pearance in the context of multi-bounce global illumination. We derive connections, where pixel and light footprints are propagated Fourier spectra for radiance and importance functions that allow us to com- independently across multiple scene interactions, in order to devise pute spatial-angular filtering footprints at path vertices, for both uni- and the first accurate bidirectional antialiasing approach. We apply our bi-directional path construction. We then use these footprints to antialias method to complex GI effects from surfaces with high-resolution reflectance modulated by high-resolution color, normal, and displacement normal, color, and displacement maps (Figures 6 to 11). maps encountered along a path. In doing so, we also unify the traditional path-space formulation of light-transport with our frequency-space inter- Our implementation is straightforward to integrate into modern pretation of global illumination pre-filtering. Our method is fully compat- renderers and we compare our filtered transport algorithms to path- ible with all existing single bounce pre-filtering appearance models, not sampling approaches with ray differentials [Igehy 1999] (when restricted by path length, and easy to implement atop existing path-space available), additionally employing different local appearance renderers. We illustrate its effectiveness on several radiometrically complex prefiltering methods (i.e., Heckbert’s diffuse color filtering [1986] scenarios where previous approaches either completely fail or require or- and Yan et al.’s specular normal map filtering [2014]).
    [Show full text]
  • Design and Development of Stream Processor and Texture Filtering Unit for Graphics Processor Architecture IP Krishna Bhushan Vutukuru1, Sanket Dessai2 1- M.Sc
    Design and Development of Stream Processor and Texture Filtering Unit for Graphics Processor Architecture IP Krishna Bhushan Vutukuru1, Sanket Dessai2 1- M.Sc. [Engg.] Student, 2- Assistant Professor Computer Engineering Dept., M. S. Ramaiah School of Advanced Studies, Bangalore 560 058. Abstract Graphical Processing Units (GPUs) have become an integral part of today’s mainstream computing systems. They are also being used as reprogrammable General Purpose GPUs (GP-GPUs) to perform complex scientific computations. Reconfigurability is an attractive approach to embedded systems allowing hardware level modification. Hence, there is a high demand for GPU designs based on reconfigurable hardware. This paper presents the architectural design, modelling and simulation of reconfigurable stream processor and texture filtering unit of a GPU. Stream processor consists of clusters of functional units which provide a bandwidth hierarchy, supporting hundreds of arithmetic units. The arithmetic cluster units are designed to exploit instruction level parallelism and subword parallelism within a cluster and data parallelism across the clusters. The texture filter unit is designed to process geometric data like vertices and convert these into pixels on the screen. This process involves number of operations, like circle and cube generation, rotator, and scaling. The texture filter unit is designed with all necessary hardware to deal with all the different filtering operations. For decreasing the area and power, a single controller is used to control data flow between clusters and between host processor and GPU. The designed architecture provides a high degree of scalability and flexibility to allow customization for unique applications. The designed stream processor and texture filtering unit are modelled in Verilog on Altera Quartus II and simulated using ModelSim tools.
    [Show full text]
  • Anti-Aliasing
    Antialiasing & Texturing Steve Rotenberg CSE168: Rendering Algorithms UCSD, Spring 2017 Texture Minification • Consider a texture mapped triangle • Assume that we point sample our texture so that we use the nearest texel to the center of the pixel to get our color • If we are far enough away from the triangle so that individual texels in the texture end up being smaller than a single pixel in the framebuffer, we run into a potential problem • If the object (or camera) moves a tiny amount, we may see drastic changes in the pixel color, as different texels will rapidly pass in front of the pixel center • This causes a flickering problem known as shimmering or buzzing • Texture buzzing is an example of aliasing Small Triangles • A similar problem happens with very small triangles • If we shoot our a single ray right through the center of a pixel, then we are essentially point sampling the image • This has the potential to miss small triangles • If we have small, moving triangles, they may cause pixels to flicker on and off as they cross the pixel centers • A related problem can be seen when very thin triangles cause pixel gaps • These are more examples of aliasing problems Stairstepping • What about the jagged right angle patterns we see at the edges of triangles? • This is known as the stairstepping problem, also affectionately known as “the jaggies” • These can be visually distracting, especially for high contrast edges near horizontal or vertical • Stairstepping is another form of aliasing Moiré Patterns • When we try to render high detail
    [Show full text]
  • Bilinear Accelerated Filter Approximation
    Eurographics Symposium on Rendering 2014 Volume 33 (2014), Number 4 Wojciech Jarosz and Pieter Peers (Guest Editors) Bilinear Accelerated Filter Approximation Paper 1019 (a) Input (b) Exact Lánczos 2 (c) Trilinear interpolation (d) CCTF (e) Our method Figure 1: Approximations of downsampling a high-resolution input image (a) to 1002 pixels using a Lánczos 2 filter are compared. The result from exact evaluation is shown in (b) and the approximations by (c) trilinear interpolation, (d) Cardinality- Constrained Texture Filtering (CCTF), and (e) our method all use the same mipmap. Trilinear interpolation appears blurry, whereas our approximation of the Lánczos 2 downsampled image is similar to the exact evaluation of a Lánczos 2 filter and CCTF. However, our method runs twice as fast as CCTF by constructing a filter from hardware accelerated bilinear texture samples. Abstract Our method approximates exact texture filtering for arbitrary scales and translations of an image while taking into account the performance characteristics of modern GPUs. Our algorithm is fast because it accesses textures with a high degree of spatial locality. Using bilinear samples guarantees that the texels we read are in a regular pattern and that we use a hardware accelerated path. We control the texel weights by manipulating the u;v parameters of each sample and the blend factor between the samples. Our method is similar in quality to Cardinality-Constrained Texture Filtering [MS13] but runs two times faster. Categories and Subject Descriptors (according to ACM CCS): I.3.3 [Computer Graphics]: Picture/Image Generation—Antialiasing 1. Introduction 3D scenes, perspective projection creates a nonuniform dis- tortion between 2D textures and the image rendered on the High-quality texture filtering is important to the appearance screen, so anisotropic filtering is commonly used to filter of a rendered image because filtering reduces aliasing arti- textures in 3D scenes.
    [Show full text]
  • Low-Resolution
    Journal of Computer Graphics Techniques Vol. 10, No. 1, 2021 http://jcgt.org Improved Shader and Texture Level of Detail Using Ray Cones Tomas Akenine-Moller¨ Cyril Crassin Jakub Boksansky NVIDIA NVIDIA NVIDIA Laurent Belcour Alexey Panteleev Oli Wright Unity Technologies NVIDIA NVIDIA Figure 1. Na¨ıve texture filtering using mipmap level 0 (left) versus our new anisotropic method (right) implemented in Minecraft with RTX on Windows 10. The insets show an 8× enlargement. During animation, the left version aliases severely, while the anisotropic ver- sion is visually alias-free. All results in this paper were generated using an NVIDIA 2080 Ti, unless otherwise mentioned. Abstract In real-time ray tracing, texture filtering is an important technique to increase image quality. Current games, such as Minecraft with RTX on Windows 10, use ray cones to determine texture-filtering footprints. In this paper, we present several improvements to the ray-cones algorithm that improve image quality and performance and make it easier to adopt in game engines. We show that the total time per frame can decrease by around 10% in a GPU-based path tracer, and we provide a public-domain implementation. 1. Introduction Texture filtering, most commonly using mipmaps [Williams 1983], is a key com- ponent in rasterization-based architectures. Key reasons for using mipmapping are to reduce texture aliasing and to increase coherence in memory accesses, making 1 ISSN 2331-7418 Journal of Computer Graphics Techniques Vol. 10, No. 1, 2021 Improved Shader and Texture Level of Detail Using Ray Cones http://jcgt.org rendering faster. Ray tracing similarly benefits from texture filtering.
    [Show full text]
  • Anisotropic Texture Filtering Using Line Integral Textures
    Online Submission ID: 0180 Anisotropic Texture Filtering using Line Integral Textures Blended Mipmap The Proposed Method 16x Anisotropic Filtering Figure 1: Our algorithm takes as input an ordinary texture with 8 bits per channel and converts it into a line integral texture with 32 bits per channel. The line integral texture allows us to achieve anisotropic filtering while using fewer samples than hardware. These images show a section of a typical game scene rendered using blended mipmapping, the line integral method, and 16x hardware anisotropic filtering. 1 Abstract 28 overblurring and/or undersampling when the texture is viewed at 29 an angle, because the necessary sampling area is anisotropic. 2 In real time applications, textured rendering is one of the most im- 3 portant algorithms to enhance realism. In this work, we propose a 30 Traditional hardware anisotropic filtering algorithms improve on 4 line integral texture filtering algorithm that is a both fast and accu- 31 these techniques by taking multiple samples inside this anisotropic 5 rate method to improving filtered image quality. In particular, we 32 sampling region. Although this approach is simple to implement, 6 present an efficient shader implementation of the algorithm that can 33 the result is merely an approximation. Also, for good results, of- 7 be compared with 16x hardware anisotropic filtering. The authors 34 tentimes many samples must be taken. For some angles, texture 8 show that the quality of the result is similar to 16x anisotropic fil- 35 quality suffers unavoidably. 9 tering, but that the line integral method requires fewer samples to 10 function, making it a more efficient rendering option.
    [Show full text]
  • Appendix C Graphics and Computing Gpus
    C APPENDIX Graphics and Computing GPUs John Nickolls Imagination is more Director of Architecture important than NVIDIA knowledge. David Kirk Chief Scientist Albert Einstein On Science, 1930s NVIDIA C.1 Introduction C-3 C.2 GPU System Architectures C-7 C.3 Programming GPUs C-12 C.4 Multithreaded Multiprocessor Architecture C-24 C.5 Parallel Memory System C-36 C.6 Floating-point Arithmetic C-41 C.7 Real Stuff: The NVIDIA GeForce 8800 C-45 C.8 Real Stuff: Mapping Applications to GPUs C-54 C.9 Fallacies and Pitfalls C-70 C.10 Concluding Remarks C-74 C.11 Historical Perspective and Further Reading C-75 C.1 Introduction Th is appendix focuses on the GPU—the ubiquitous graphics processing unit graphics processing in every PC, laptop, desktop computer, and workstation. In its most basic form, unit (GPU) A processor the GPU generates 2D and 3D graphics, images, and video that enable window- optimized for 2D and 3D based operating systems, graphical user interfaces, video games, visual imaging graphics, video, visual computing, and display. applications, and video. Th e modern GPU that we describe here is a highly parallel, highly multithreaded multiprocessor optimized for visual computing. To provide visual computing real-time visual interaction with computed objects via graphics, images, and video, A mix of graphics the GPU has a unifi ed graphics and computing architecture that serves as both a processing and computing programmable graphics processor and a scalable parallel computing platform. PCs that lets you visually interact with computed and game consoles combine a GPU with a CPU to form heterogeneous systems.
    [Show full text]
  • Sampling II, Anti-Aliasing, Texture Filtering
    ME-C3100 Computer Graphics, Fall 2019 Jaakko Lehtinen Many slides from Frédo Durand Sampling II, Anti-aliasing, Texture Filtering CS-C3100 Fall 2019 – Lehtinen 1 Admin and Other Things • Give class feedback – please! – You will get an automated email with link – I take feedback seriously – One additional assignment point for submitting • 2nd midterm / exam on Wed Dec 18 – Two sheets of two-sided A4 papers allowed – Those who took 1st midterm: only complete 1st half – Those who didn’t: answer all questions • Last lecture today Aalto CS-C3100 Fall 2019 – Lehtinen 2 Sampling • The process of mapping a function defined on a continuous domain to a discrete one is called sampling • The process of mapping a continuous variable to a discrete one is called quantization • To represent or render an image using a computer, we must both sample and quantize – Today we focus on the effects of sampling continuous discrete value value continuous position discrete position CS-C3100 Fall 2019 – Lehtinen 3 Sampling & Reconstruction The visual array of light is a continuous function 1/ We sample it... – with a digital camera, or with our ray tracer – This gives us a finite set of numbers (the samples) – We are now in a discrete world 2/ We need to get this back to the physical world: we reconstruct a continuous function from the samples – For example, the point spread of a pixel on a CRT or LCD – Reconstruction also happens inside the rendering process (e.g. texture mapping) • Both steps can create problems – Insufficient sampling or pre-aliasing (our focus
    [Show full text]